1. Ryan Purdy, FCAS, MAAA Merlinos & Associates
Price Optimization
Ryan Purdy, FCAS, MAAA
Merlinos & Associates

2. GET READY FOR EXCITEMENT!

3. WHAT WE WILL TALK ABOUT…
Introduce You to my Friends
How Actuaries Optimize Price
Optimized Prices are not Price Optimization
Recap and Regulatory Concerns

4. How Do We Explain Price Optimization?
Karen

5. I have a few friends…
Karen
Bob
Donald

6. I have lots of friends…

7. All my friends owe me money…
…because I sell them insurance!

8. BUT HOW MUCH MONEY DO THEY OWE ME?
Fundamental question of actuarial science.
How do I collect money in a fair way as to make money?
If I don’t do this correctly, I will lose some friends and lose money.

9. BUT HOW MUCH MONEY DO THEY OWE ME?
Actuaries have always tried to optimize prices to promote long-term stability in the insurance markets.

10. BUT HOW MUCH MONEY DO THEY OWE ME?
This is codified in our Statements of Principles, which state:
“A rate is reasonable and not excessive, inadequate, or unfairly discriminatory if it is an actuarially-sound estimate of the expected value of all future costs associated with an individual risk transfer!”

11. BUT HOW MUCH MONEY DO THEY OWE ME?
So how do we go about optimizing price?
Let’s return to our friends to discuss.

12. First, how much money do I need from everyone combined…
$20 Million

13. First, how much money do I need from everyone combined…
$20 Million
This doesn’t require I know anything about the individuals, just how the group has performed.
This is what an actuarial indication is doing.

14. THERE ARE DIFFERENCES
But there are differences in each one of my policyholders…
Some are more risky, others are lower
If I fail to recognize this, and charge equivalent amounts to everyone…
I will lose less risky folks over time because I am over charging them.

15. THERE ARE DIFFERENCES
Traditional risk classification finds groups of characteristics that are related to loss potential.
We would look at experience groups of risks:
Expensive Cars vs Cheaper Cars
Old Cars vs New Cars
Males vs Females, Married vs Single

16. Risk Classification recognizes shades of risks…
$20 Million

17. Risk Classification recognizes shades of risks…
$20 Million
Risk Classification tells us how to charge different groups.
I still need the same amount of money overall.

18. MODERN WAYS TO OPTIMIZE PRICES
Rise of analytics has allowed for new ways to analyze pricing.
EXAMPLES
Predictive Analytics
CAT Model Analytics
Reinsurance Allocation Analytics

19. Predictive Analytics What is Predictive Analytics…
Business analysis that produces a predictive score for each customer, prospect, claim, etc. 65 11 48 53 16 27 33 84

20. Predictive Analytics
When used in pricing, predictive analytics helps us understand how likely each individual is to have a claim and how severe that claim might be.
In contrast, traditional actuarial classification analysis tells us the same thing for groups of individuals.

21. Predictive Analytics
Use of predictive analytics has allowed for highly granular pricing systems to develop!
Helped develop and support many new rating variables now common in the industry like credit.
Use has led to pricing manuals that now contain thousands of pages and billions of possible price points.

22. Catastrophe Model Analytics
Cat Models aren’t just for risk management anymore!!!

23. Catastrophe Model Analytics
Modern pricing plans are using cat models to directly produce pricing!
We can develop granular pricing for variables that would not be possible using actual data!

24. Catastrophe Model Analytics
We can now create pricing for the individual risk for the catastrophe perils.
EXAMPLES
Wind mitigation features
Distance-to-Coastlines
Distance-to-Urban/Rural Interface
Soil Liquefaction Potential at Property Location

25. Reinsurance Allocation Analytics
Cat Models are also allowing complex reinsurance allocations!

26. Reinsurance Allocation Analytics
Layer 3
Layer 2
Layer 1

27. Reinsurance Allocation Analytics

28. Reinsurance Allocation Analytics
Layer 1

29. Reinsurance Allocation Analytics
Detailed model information can let us find how each member of the portfolio contributes to the need to buy reinsurance, and make proper cost allocations to those risks.
Analysis is logical, but can require analysis of billions of lines of data.
Most advanced carriers are trying to produce real-time pricing at quote based on portfolio detail.

30. MODERN WAYS TO OPTIMIZE PRICES
These examples all strive to more accurately match cost to price at the policy level!
Let’s revisit our friends!

31. Modern pricing adds color to our risk assessments…
$20 Million

32. Modern pricing adds color to our risk assessments…
These techniques can give us unique pricing for each customer.
Again, I still need the same amount of money overall. Just need it differently from different folks.

33. So That is How we Optimize Prices!
Using modern techniques we can generate optimized prices that reflect the costs of individual risk transfer:
“A rate is reasonable and not excessive, inadequate, or unfairly discriminatory if it is an actuarially sound estimate of the expected value of all future costs associated with an individual risk transfer!”

34. So That is How we Optimize Prices!
None of this is Price Optimization!

35. Sell Karen auto coverage at the most accurate estimate of cost
Optimized Prices!
Sell Karen auto coverage at the most accurate estimate of cost

36. I have a customer relationship with Karen that has value!
Price Optimization
I have a customer relationship with Karen that has value!

37. Price Optimization
Change in focus of pricing from individual risk transfer to focus on providing optimal relationship value!
…each relationship is different
…each consumer is different
…each has a different value

38. Price Optimization – Step 1
First step is the same…
I still need to find the most accurate estimate of costs for all my potential product lines!

39. Price Optimization – Step 1
$20 Million

40. Price Optimization – Step 2
$20 Million

41. Price Optimization – Step 2
Year 1
Year 2
Year 3
Year 4
Year 5

42. Price Optimization – Step 2
Year 1
Year 2
Year 3
Year 4
Year 5
Retention modeling predicts how long each customer might be with me.
This might differ product by product.
5 Year folks are higher value than 1 Year folks

43. Price Optimization – Step 3
Year 1
Year 2
Year 3
Year 4
Year 5

44. Price Optimization – Step 3
Year 1
Year 2
Year 3
Year 4
Year 5

45. Price Optimization – Step 3
Year 1
Year 2
Year 3
Year 4
Year 5

46. Price Optimization – Step 3
Year 1
Year 2
Year 3
Year 4
Year 5

47. Price Optimization – Step 3
Year 1
Year 2
Year 3
Year 4
Year 5
Retention and cross-selling modeling predicts customer value across product offerings.
Longer term customers with multiple product lines more value than mono-line

48. Price Optimization – Step 4
Year 1
Year 2
Year 3
Year 4
Year 5

49. Price Optimization – Step 4
Profit = $100
Cost
Profit = $20
Cost
Profit = $80
Cost

50. Price Optimization – Step 4
Processes for price optimization up to this point have simply recognized that some customer relationships are:
Longer lasting
Multiple
More Valuable Long-term
This all seems reasonable, what is the issue?

51. Price Optimization – Step 5
Main point of contention is a concept called the price elasticity of demand!
In short, this concept recognizes that individual consumers have different tolerances for prices and pricing changes in the market.

52. Price Optimization – Step 5
By modeling this behavior, I can figure out how customers will react to price changes:
Buy more coverage at cheaper rates?
Leave/drop coverage at higher prices?
Stay longer/Leave sooner?
Opposite and unexpected results?
Figures this out customer by customer.

53. Price Optimization – Step 5
Year 1
Year 2
Year 3
Year 4
Year 5
Base Case
10% Reduction
15% Increase
5% Increase
25% Increase

54. Price Optimization – Step 5
Profit = $70
Base Case
Cost
10% Reduction
Profit = $80
Cost
15% Increase
Cost
Profit = $50
Profit = $50
5% Increase
Cost
25% Increase
Cost
Profit = $100

55. Price Optimization – Step 5
$27 Million

56. Price Optimization
When this process is extended across all consumers, I optimize my profit across a specified horizon.
Our prior example was simple and logical:
What happens if some people don’t shop, even when facing large rate increases?
Price Optimization models would suggest increase as much as possible.

57. Price Optimization
Regulators contend this process is not actuarially sound, and against statutes:
“A rate is reasonable and not excessive, inadequate, or unfairly discriminatory if it is an actuarially sound estimate of the expected value of all future costs associated with an individual risk transfer!”

58. Summary
Price Optimization can depart from true cost-based pricing and turn into profit optimization.
Practice is common-place in European markets, and U.S. carriers are starting to explore!
Proponents argue that they have modified routines so that no one gets an unsound rate, but transparency is lacking

59. Summary
Most company’s still have a lot of areas they can improve cost-based pricing.
Additional scrutiny can be expected for anything that is non-traditional.
It is important to be able to explain pricing analytics in clear ways to be able to navigate the heightened regulatory scrutiny.

60. Price Optimization: Industry Evolution
RYAN PURDY, FCAS, MAAA
MERLINOS & ASSOCIATES
Thank you!